JPWO2013061664A1 - Ultrasonic imaging apparatus, ultrasonic imaging method, and ultrasonic imaging program - Google Patents

Abstract

Provided is an ultrasonic imaging apparatus capable of directly creating a scalar field image and grasping a tissue boundary without obtaining a motion vector. The received signal is processed to generate an image of two or more frames, two frames are selected from the images, a plurality of regions of interest in one frame, and a plurality of regions of interest that are wider than the region of interest in another frame Set for each. A plurality of candidate regions having a size corresponding to the region of interest are set in the search region. The norm between the pixel value in the region of interest and the pixel value in the candidate region is obtained for each of the plurality of candidate regions, thereby obtaining the norm distribution in the search region. An image is generated using a value representing the norm distribution state as a pixel value of a region of interest corresponding to the search region.

Description

  The present invention relates to an ultrasonic imaging method and an ultrasonic imaging apparatus that can clearly identify a tissue boundary when imaging a living body using ultrasonic waves.

  In an ultrasound imaging device used for medical image diagnosis, the elasticity coefficient distribution of a tissue is estimated based on the amount of change in a small area of a diagnostic moving image (B-mode image), and the hardness is converted into a color map for display. How to do is known. However, for example, in the case of the peripheral part of a tumor, the acoustic impedance and the elastic modulus may not differ greatly with respect to the surrounding tissue. I cannot grasp the boundary with other organizations.

  Therefore, there is a method of obtaining a motion vector of each region of a diagnostic image by block matching processing for two different diagnostic image data in time series and creating a scalar field image from the motion vector. As a result, it is possible to identify a boundary between tissues whose acoustic impedance and elastic modulus are not significantly different from those of the surroundings.

  However, in areas where there is a lot of noise in the image data, such as a penetration limit area where the echo signal is weak, an error vector is generated due to the influence of noise when obtaining a motion vector, and the boundary discrimination is degraded. Therefore, in Patent Document 1, when obtaining a motion vector, the similarity of image data between a plurality of regions that are candidates for the destination of the region of interest and the region of interest is obtained, and the motion obtained for the region of interest from the similarity distribution The reliability of the vector is judged. When the reliability of the motion vector is low, the motion vector can be removed and the boundary identification can be improved.

International Publication WO2011 / 052602

  A conventional method of identifying a tissue boundary by obtaining a motion vector described in Patent Document 1 or the like obtains a motion vector of each region on an image by block matching processing, converts the motion vector into a scalar, and generates a scalar field image. Two steps are required:

  An object of the present invention is to provide an ultrasonic imaging apparatus that can directly create a scalar field image and grasp the boundary of an object without obtaining a motion vector.

  In order to achieve the above object, according to the first aspect of the present invention, the following ultrasonic imaging apparatus is provided. That is, a transmission unit that transmits ultrasonic waves toward a target, a reception unit that receives ultrasonic waves coming from the target, and a processing unit that processes a reception signal of the reception unit to generate an image of two or more frames. . The processing unit sets a plurality of regions of interest in one frame among the generated images of two or more frames, and sets a search region wider than the region of interest for each of the plurality of regions of interest in another one frame. , By setting a plurality of candidate regions of a size corresponding to the region of interest in the search region, and obtaining a norm between the pixel value of the region of interest and the pixel value in the candidate region for each of the plurality of candidate regions, A norm distribution in the search region is obtained, and an image is generated using a value (scalar value) representing the norm distribution state as a pixel value of the region of interest corresponding to the search region.

  According to the present invention, a value representing the norm distribution state of the search area is obtained. The norm indicates a low value along the boundary if there is a boundary. Therefore, since an image is generated using a value representing the norm distribution state (scalar value) as a pixel value of the region of interest corresponding to the search region, an image representing the boundary of the subject is generated without generating a vector field. be able to.

1 is a block diagram showing a system configuration example of an ultrasonic imaging apparatus according to a first embodiment. 5 is a flowchart showing a processing procedure for image generation by the ultrasonic imaging apparatus according to the first embodiment. The flowchart which shows the detail of step 24 of FIG. The figure explaining the process of step 24 of FIG. 2 with the test object (phantom) of 2 layer structure. (A) is a distribution diagram showing the p-norm distribution of the search region when the region of interest is in the stationary part, (b) is a histogram of the p-norm distribution of (a), and (c) is the first implementation. FIG. 7D is a distribution diagram showing the p-norm distribution of the search area when the region of interest is at the boundary, and FIG. 6D is a histogram of the p-norm distribution of FIG. (A) B-mode image of the first embodiment, (b) Scalar field image of the first embodiment, (c) conventional vector field image, (d) distortion tensor image of the conventional vector field image. (A) Explanatory drawing which shows the superimposed image of the scalar field image and B mode image of 1st Embodiment, (b) The scalar field image of 1st Embodiment, B mode image, and a vector field image Explanatory drawing which shows a superimposed image. 5 is a flowchart showing a processing procedure for image generation by the ultrasonic imaging apparatus according to the first embodiment. (A) Image showing an example in which a virtual image is generated in a scalar field image, (b) Histogram of average value and frequency of p-norm value of second embodiment, (c) Dark color of low-reliability part in second embodiment Scalar field image replaced with display. 9 is a flowchart illustrating an image processing procedure according to the second embodiment. 10 is a flowchart illustrating an image generation procedure according to the third embodiment. (A)-(h) Explanatory drawing which shows the model of eight directions set to the search area | region of 3rd Embodiment. (A)-(c) Explanatory drawing which shows the direction of the boundary of 6th Embodiment, and the example of a pattern of a vector field. The flowchart which shows the procedure which calculates | requires the boundary norm of 6th Embodiment. Explanatory drawing which shows ROI set partially overlapping in 7th Embodiment. The flowchart which shows the process sequence for the amount of calculations reduction using the look-up table of 7th Embodiment. The graph which shows the information entropy calculated | required about the continuous frame in 8th Embodiment. The flowchart which shows the process sequence of the image display using the information entropy of 8th Embodiment. (A) Explanatory drawing which shows the superimposed image of the extracted scalar field image of 9th Embodiment, a vector field image, and a B-mode image, (b) The histogram of the scalar value and frequency of a scalar field image. The flowchart which shows the process sequence which produces | generates the extracted scalar field image of 9th Embodiment.

  An ultrasonic imaging apparatus according to the present invention includes a transmission unit that transmits ultrasonic waves toward a target, a reception unit that receives ultrasonic waves coming from the target, and processes a reception signal of the reception unit to generate an image of two or more frames. And a processing unit to be generated. The processing unit sets a plurality of regions of interest in one frame among the generated two or more frames, and sets a search region wider than the region of interest in another frame for each of the plurality of regions of interest. A plurality of candidate regions having a size corresponding to the region of interest are set in the search region. The processing unit obtains a norm between the pixel value in the region of interest and the pixel value in the candidate region for each of the plurality of candidate regions, thereby obtaining a norm distribution in the search region, and representing the norm distribution state An image is generated using (scalar value) as the pixel value of the region of interest corresponding to the search region. Here, in the region of interest, it is also possible to calculate the norm by directly using the amplitude value or phase value of the received signal instead of the pixel value. Since the pixel value is logarithmically compressed, a linear change is more accurately reflected in the original received signal, and high resolution is achieved.

  Since the norm indicates a low value along the boundary if there is a boundary, a value representing the distribution state of the norm (scalar value) is generated as a pixel value of the region of interest corresponding to the search region. The ultrasonic imaging apparatus of the invention can generate an image representing the boundary of the subject without generating a vector field.

  As the norm, a p-norm (also referred to as a power norm) represented by the following formula (1) can be used.

ただし、P_m(i_０、j_０)は、前記関心領域内の所定の位置(i_０、j_０)（例えば中心位置）にある画素の画素値、P_m＋Δ(i、j)は、前記候補領域内の位置(i、j)にある画素の画素値、ｐは予め定めた実数である。

However, P _m (i ₀ , j ₀ ) is a pixel value of a pixel at a predetermined position (i ₀ , j ₀ ) (for example, the center position) in the region of interest, and P _{m + Δ} (i, j) is The pixel value of the pixel at position (i, j) in the candidate area, p is a predetermined real number.

  The p is preferably a real number larger than 1.

  As a value (scalar value) representing the norm distribution state, for example, a norm distribution statistic is used. For example, as the statistic, a divergence degree defined by the difference between the minimum value of the norm value of the norm distribution in the search region and the average value of the norm values can be used. Further, for example, a variation coefficient obtained by dividing the standard deviation of the norm value of the norm distribution in the search region by the average value can be used as the statistic.

  As a value (scalar value) representing the norm distribution state, a value other than a statistic can be used. For example, out of a plurality of directions centered on the region of interest set in the search region, the first direction in which the average norm value of candidate regions at positions along the direction is minimized, and the first direction through the region of interest A second direction orthogonal to one direction is obtained, and a ratio value or a difference value between an average norm value of candidate areas along the first direction and an average norm value of candidate areas along the second direction is obtained. It can be used as a value representing the norm distribution state for the region of interest corresponding to the search region. At this time, the norm distribution in the search region may be enhanced by a Laplacian filter in advance, and a ratio value or a difference value may be obtained for the distribution after the enhancement process.

  In addition, a matrix representing the norm distribution in the search area is generated, eigenvalue decomposition processing is performed on the matrix to obtain an eigenvalue, and this eigenvalue is a value (scalar) representing the norm distribution state for the region of interest corresponding to the search area. Value).

  Further, the processing unit may be configured to further obtain a motion vector. For example, the processing unit selects a candidate region having a minimum norm value in the search region as a destination of the region of interest, and obtains a motion vector that connects the position of the region of interest and the position of the selected candidate region. A motion vector field is generated by generating a motion vector for each of a plurality of regions of interest. Further, the processing unit sets the sum of the square value of the y-direction derivative with respect to the x component and the square value of the x-direction derivative with respect to the y component as a boundary norm value for each of the plurality of regions of interest set in the motion vector field. It is also possible to generate an image using this boundary norm value as the pixel value of the region of interest.

  In addition, when the processing unit is set to partially overlap a plurality of regions of interest, the value obtained for the overlapping region when calculating the norm for one region of interest is stored in the lookup table of the storage region. It is also possible to store and read out from the look-up table when calculating the norm for other regions of interest. Similarly, when a plurality of candidate areas are set to partially overlap, when the values obtained for the overlapping areas are stored in the lookup table of the storage area and the norm is calculated for the other candidate areas It is also possible to read out from the lookup table and use it. As a result, the amount of calculation can be reduced.

  In addition, the processing unit generates a plurality of frames in a time series of images generated by using values representing the norm distribution state as pixel values, calculates an information entropy amount for each frame, and the information entropy amount is greater than a preset threshold value. If it is small, it is possible not to use it as an image for displaying a frame. Thereby, since an abnormal image with a small information entropy amount can be eliminated, a continuous image with good visibility can be displayed.

  It is also possible to generate an extracted image obtained by extracting pixels having a norm distribution value equal to or greater than a predetermined value and display the extracted image superimposed on the B-mode image. A pixel whose value representing the norm distribution state is equal to or greater than a predetermined value is a pixel indicating a boundary portion, so that an extracted image can be displayed only at the boundary portion of the B-mode image. In order to determine the predetermined value, for an image generated with a pixel value representing a norm distribution state, a value representing the norm distribution state and a histogram of the frequency thereof are generated, and a histogram-like distribution of the histogram is searched. It is also possible to use the minimum value of the mountain-shaped distribution as the predetermined value.

  In addition, an ultrasonic imaging apparatus according to another aspect of the present invention includes a transmission unit that transmits ultrasonic waves toward a target, a reception unit that receives ultrasonic waves coming from the target, and a reception signal of the reception unit And a processing unit that generates an image of two or more frames. The processing unit sets a plurality of regions of interest in the received signal distribution corresponding to one frame among the received signals corresponding to the received images of two or more frames, and is interested in the received signal distribution corresponding to another one frame. A search area wider than the area is set for each of the plurality of regions of interest, a plurality of candidate regions corresponding to the region of interest are set in the search region, the amplitude distribution or phase distribution of the region of interest, and the amplitude in the candidate region By calculating the norm between the distribution and the phase distribution for each of the plurality of candidate regions, the norm distribution in the search region is obtained, and the value representing the distribution state of the norm is set as the pixel of the region of interest corresponding to the search region. Generate an image as a value.

  In addition, according to the present invention, an ultrasonic imaging method is provided. That is, an ultrasonic wave is transmitted toward the target, and a reception signal obtained by receiving the ultrasonic wave coming from the target is processed to generate an image of two frames or more. Two frames are selected from the image, a plurality of regions of interest are set in one frame, and a search region wider than the region of interest is set in another frame for each region of interest. A plurality of candidate regions having a size corresponding to the region of interest are set in the search region. The norm between the pixel value in the region of interest and the pixel value in the candidate region is obtained for each of the plurality of candidate regions, thereby obtaining the norm distribution in the search region. An image is generated using a value representing the norm distribution state as a pixel value of the region of interest corresponding to the search region.

  According to the present invention, an ultrasound imaging program is provided. That is, a first step of selecting two frames from an ultrasonic image of two or more frames in a computer, a second step of setting a plurality of regions of interest in one frame, and the region of interest in another one frame A third step of setting a wide search area for each of a plurality of regions of interest, and setting a plurality of candidate regions of a size corresponding to the region of interest in the search region, a pixel value of the region of interest and the pixels in the candidate region A fourth step for obtaining a norm distribution in the search region by obtaining a norm between the values for each of the plurality of candidate regions, and a value representing the distribution state of the norm in the region of interest corresponding to the search region It is the program for ultrasonic imaging for performing the 5th step which produces | generates an image as a pixel value.

  An ultrasonic imaging apparatus according to an embodiment of the present invention will be specifically described below.

(First embodiment)
FIG. 1 shows a system configuration of the ultrasonic imaging apparatus of the present embodiment. This apparatus has an ultrasonic boundary detection function. As shown in FIG. 1, this apparatus includes an ultrasonic probe (probe) 1, a user interface 2, a transmission beam former 3, a control system 4, a transmission / reception changeover switch 5, a reception beam former 6, and an envelope detection unit 7. , A scan converter 8, a processing unit 10, a parameter setting unit 11, a combining unit 12, and a display unit 13.

  An ultrasonic probe 1 in which ultrasonic elements are arranged one-dimensionally is a transmitter that transmits an ultrasonic beam (ultrasonic pulse) to a living body, and an echo signal (received signal) reflected from the living body. Is a receiving unit. Under the control of the control system 4, a transmission signal having a delay time adjusted to the transmission focus is output by the transmission beamformer 3 and sent to the ultrasonic probe 1 via the transmission / reception changeover switch 5. The ultrasonic beam reflected or scattered in the living body and returned to the ultrasonic probe 1 is converted into an electric signal by the ultrasonic probe 1 and received by the receiving beam former 6 via the transmission / reception switch 5. Sent as.

  The receive beamformer 6 is a complex beamformer that mixes two received signals that are 90 degrees out of phase, and performs dynamic focus that adjusts the delay time according to the reception timing under the control of the control system 4. Outputs real and imaginary RF signals. The RF signal is detected by the envelope detector 7 and then converted into a video signal, which is input to the scan converter 8 and converted into image data (B-mode image data). The configuration described above is the same as the configuration of a known ultrasonic imaging apparatus. Furthermore, in the present invention, it is also possible to perform ultrasonic boundary detection by a configuration that directly processes the RF signal.

  In the apparatus of the present invention, the ultrasonic boundary detection process is realized by the processing unit 10. The processing unit 10 includes a CPU 10a and a memory 10b. The CPU 10a executes a program stored in the memory 10b in advance, thereby generating a scalar field image capable of detecting the boundary of the subject tissue. The scalar field image generation processing will be described in detail later with reference to FIG. The combining unit 12 combines the scalar field image and the B-mode image and then displays them on the display unit 13.

  In the parameter setting unit 11, parameters for signal processing in the processing unit 10, selection setting of a display image in the synthesis unit 12, and the like are performed. These parameters are input from the user interface 2 by an operator (device operator). As parameters for signal processing, for example, setting of a region of interest on a desired frame m and setting of a search region on a frame m + Δ different from the frame m can be received from an operator. As the selection setting of the display image, for example, the selection setting of whether the original image and the vector field image (or scalar image) are combined into one image and displayed on the display, or two or more moving images are displayed side by side. Can be received from the operator.

  FIG. 2 is a flowchart showing the operation of image generation and composition processing in the processing unit 10 and the composition unit 12 of the present invention. The processing unit 10 first acquires a measurement signal from the scan converter 8 and performs normal signal processing to create a B-mode moving image (steps 21 and 22). Next, a desired frame m and two frames m + Δ at different times are extracted from the B-mode moving image (step 23). For example, Δ = 1 frame, and the desired frame m and the next frame m + 1 are extracted. The extraction of the two frames can be received from the operator via the parameter setting unit 11 or can be configured to be automatically selected by the processing unit 10.

  The processing unit 10 calculates a p-norm distribution from the two extracted frames and generates a scalar field image (step 24). A composite image in which the obtained scalar field image is superimposed on the B-mode image is generated and displayed on the display unit 13 (step 27). In step 23, it is also possible to display a moving image of the synthesized image by selecting different frames sequentially in time series as the desired frame, repeating the above steps 21 to 27, and continuously displaying the synthesized image.

FIG. 3 is a flowchart showing detailed processing of the scalar field image generation operation in step 24 described above. First, the processing unit 10 sets a region of interest (ROI) 31 having a predetermined number of pixels N in the frame m extracted in step 23 as shown in FIG. 4 (step 51). A pixel value of a pixel included in the ROI 31, for example, a luminance distribution is represented as P _m (i ₀ , j ₀ ). i ₀ and j ₀ indicate the position of the pixel in the ROI 31.

  Next, the processing unit 10 sets a search area 32 of a predetermined size in the frame m + Δ extracted in step 23 as shown in FIG. 4 (step 52). The search area 32 includes the position of the ROI 31 of the frame m. For example, the search area 32 is set to coincide with the center position of the ROI 31. The size of the search area 32 is a predetermined size larger than the ROI 31. Here, a configuration in which the ROI 31 is sequentially set over the entire image of the frame m and the search area 32 having a predetermined size is set centered on the ROI 31 will be described, but the frame m received from the operator in the parameter setting unit 11 will be described. It is also possible to sequentially set the ROI 31 only in the predetermined range.

The processing unit 10 sets a plurality of candidate areas 33 having the same size as the ROI 31 as shown in FIG. In FIG. 4, the search area 32 is divided into a matrix having the same size as the ROI 31 and the candidate area 33 is set. However, the adjacent candidate areas 33 may be set so as to partially overlap each other. A pixel value of a pixel included in the candidate area 33, for example, a luminance distribution is represented as P _{m + Δ} (i, j). i and j indicate the positions of the pixels in the candidate area 33.

The processing unit 10 uses the luminance distribution P _{m + Δ} (i, j) of the pixels in the candidate region 33 and the luminance distribution P _m (i ₀ , j ₀ ) of the ROI 31 to calculate the p-norm by the above-described equation (1). This is calculated and used as the p-norm value of the candidate area 33. P-norm of the above-mentioned equation (1) is the position of ROI31 _{(i 0,} j ₀₎ luminance _{P m (i 0, j 0} ) of pixels, corresponding to the position _{i 0,} j ₀ in the candidate area 33 The absolute value of the difference from the luminance P _{m + Δ} (i, j) of the pixel at the position (i, j) is obtained to the p-th power, added for all the pixels in the candidate region 33, and the 1 / p-th power. As the p value, a predetermined real number or a value received from the operator via the parameter setting unit 11 is used. The p value is not limited to an integer, and may be a decimal number.

The p-norm where p is a power number as in the above formula (1) is a value corresponding to the concept of distance, and the luminance distribution P _m (i ₀ , j ₀ ) of the ROI 31 and the luminance distribution of the candidate region 33. The similarity of P _{m + Δ} (i, j) is shown. That is, if the luminance distribution P _m (i ₀ , j ₀ ) of the ROI 31 and the luminance distribution P _{m + Δ} (i, j) of the candidate region 33 are the same, the p-norm is zero. Moreover, the larger the brightness distribution of the two, the larger the value.

  The processing unit 10 obtains a p-norm value for all candidate regions 33 of the search region 32 (step 53). Thereby, the p-norm distribution in the search region 32 corresponding to the ROI 31 can be obtained. The obtained p-norm value is stored in the memory 10b in the processing unit 10.

  5A and 5C show examples of the p-norm distribution of the present invention. FIG. 5A shows the norm distribution of the search region 32 when the ROI 31 and its search region 32 are both located in the stationary part of the subject. FIG. 5C shows an object in which the phantom 41 and the phantom 42, which are gel base materials, are vertically stacked, and the ROI 31 is placed on the boundary where the lower phantom slides relative to the upper phantom 41 in the horizontal direction. This is the norm distribution of the search region 32 when arranged. 5A and 5C, the search area 32 is divided into 21 × 21 candidate areas 33. The block size of the candidate area 33 is 30 × 30 pixels, the search area 32 is 50 × 50 pixels, and the candidate area 33 is moved pixel by pixel within the search area 32. That is, the candidate areas 33 adjacent to each other overlap 29 pixels. The center of the search area 32 corresponds to the position of the ROI 31.

  As shown in FIG. 5A, when the ROI 31 is located in the stationary part, the center position corresponding to the position of the ROI 31 has a minimum norm value in the p-norm distribution. On the other hand, as shown in FIG. 5C, when the ROI 31 is arranged at the boundary of the sliding subject, the norm distribution is not only the center position of the search region 32 becomes the minimum norm value, but also within the search region 32. A region having a small p-norm value (valley of p-norm) is formed in a direction along the boundary of the subject.

  In this way, the present invention uses and images the fact that the distribution of the p-norm of the search region 32 differs depending on whether the ROI 31 is located at the stationary part of the subject or at the sliding boundary. Specifically, a statistic indicating the p-norm distribution of the search area 32 is obtained and set as a scalar value of the ROI 31 corresponding to the search area 32 (step 54). Any statistics can be used as long as it can represent the difference in the p-norm distribution between the stationary part and the boundary part. Here, the divergence degree calculated | required by Formula (2) is used as a statistic.

すなわち、処理部１０は、探索領域３２の全ｐノルム値から、最小値と平均値を求め、式（２）により乖離度を演算する。

That is, the processing unit 10 obtains a minimum value and an average value from all the p-norm values in the search region 32, and calculates the divergence degree using Expression (2).

  The divergence degree of Expression (2) is shown using the histograms of FIGS. 5 (b) and 5 (d). FIGS. 5B and 5D are histograms of p-norm values in the search region 32 in FIGS. 5A and 5C, respectively. When the ROI 31 is located in a stationary part as shown in FIG. 5A, the p-norm distribution of the search region 32 has a minimum norm value at the center position corresponding to the position of the ROI 31, and the p-norm value around it is high. It can be seen that the minimum value of the p-norm value is sufficiently different from the average of the distribution. Therefore, the degree of divergence increases. On the other hand, when the ROI 31 is located at the boundary as shown in FIG. 5C, the p-norm value at the center position corresponding to the ROI 31 becomes the minimum value, but the p-norm value also decreases in the region along the surrounding boundary. Since there is an error, the distribution of the entire histogram is widened. For this reason, the difference between the minimum value of the p-norm and the average of the distribution is reduced, and the degree of deviation is also reduced.

  In this way, in step 54, by obtaining the divergence degree (scalar value) of the p-norm distribution, it is indicated by the scalar value whether the ROI 31 is located at the stationary part of the subject or at the sliding boundary. Can do.

  The above steps 51 to 54 are repeated until all ROIs 31 are calculated (step 55). An image (scalar field image) is generated by converting the degree of divergence (scalar value) obtained for all ROIs 31 into pixel values (for example, luminance values) of the image (step 56). Through the above steps 51 to 56, the scalar field image of step 24 is generated.

  FIG. 6A shows a specific example of the B-mode image obtained at step 22 and FIG. 6B shows a specific example of the scalar field image obtained at step 24. The B-mode image in FIG. 6A is obtained by superposing the gel base material phantoms 41 and 42 in two layers and fixing the ultrasonic probe to the upper phantom 41 and moving it laterally. The upper side where the ultrasonic probe is fixed is relatively stationary, while the lower phantom 42 is a vector field indicating lateral movement. In FIG. 6B, the calculation is performed with p = 2.

  As shown in FIG. 6 (b), the scalar field image with the divergence degree (scalar value) of the p-norm distribution of the present invention as the pixel value has a large divergence degree at the sliding boundary part of the phantoms 41 and 42, and the sliding boundary part is It can be seen that it is clearly imaged. Therefore, in step 26, by displaying the scalar field image alone or superimposed on the B-mode image, the boundary where the boundary is difficult to appear in the B-mode image, for example, the boundary of the tissue whose acoustic impedance and elastic modulus are not significantly different from the surroundings. , Can be clearly displayed in a scalar field image.

  Further, in the scalar field image of FIG. 6B, it can be seen that a virtual image is not generated even in a deep region near the penetration limit, and the virtual image is suppressed.

  As a comparative example, FIG. 6C shows a vector field image obtained by a conventional block matching process. In this image, the B-mode image in FIG. 6A is set as the frame m, the ROI moving position is obtained by block matching processing (step 24) with the next frame (frame m + Δ, Δ = 1 frame), and the moving direction is determined. It is the vector field image which showed the magnitude | size as a vector (arrow). In FIG. 6C, a phenomenon in which the motion vector is disturbed is seen in the lower (deep) region. This is because the SN ratio of the detection sensitivity decreases and the influence of electric noise and the like increases as the distance from the probe installed at the upper portion increases, indicating the penetration limit. FIG. 6D shows a distortion tensor of the vector field shown in FIG. 6C, which is imaged as a pixel value (for example, a luminance value). In FIG. 6C as well, a phenomenon in which a virtual image is generated in the deep part due to the influence of an erroneous vector can be seen.

  Thus, the tensor field image (FIG. 6B) obtained from the p-norm of the present embodiment is compared with the conventional vector field of FIG. 6C and FIG. 6D and the distortion tensor field image based on the vector field. Thus, the virtual image can be suppressed and the slip boundary portion can be clearly shown.

  The scalar field image and the B-mode image obtained in the present invention are displayed in a superimposed manner as shown in FIG. Thereby, even when the boundary of the B-mode image is unclear, the boundary can be grasped by the scalar field image.

In the present embodiment, it is also possible to further generate a vector field image and display the vector field image superimposed on the scalar field image and the B-mode image. In this case, step 25 is performed after step 24 as shown in FIG. In step 25, the minimum value is searched for among the p-norm values of all candidate areas 33 in the search area 32 calculated in step 24, and the candidate area 33 having the minimum value is determined as the destination area of the ROI 31. A motion vector that connects the position (i ₀ , j ₀ ) and the position (i _min , j _min ) of the candidate area to be moved is determined. By executing this for all the ROIs 31, a vector field is obtained. An image in which each vector is indicated by an arrow can be generated to obtain a vector field image.

  The obtained vector field image is displayed superimposed on the scalar field image and the B-mode image (step 26). An example of the superimposed image is shown in FIG. By superimposing and displaying the vector field image, it is possible to clearly grasp in which direction the subject is moving across the boundary confirmed in the scalar field image.

  In the present embodiment, the p-value of the p-norm of Equation (1) may be a real number, but the optimal p value is obtained by conducting a parameter survey of the p-value with an appropriate change width, for example, for a typical sample to be evaluated. It is possible to set the value to obtain a clear image with the fewest virtual images. The p value is more preferably a real number larger than 1. FIG. 9A shows an image example when a scalar field image is obtained by the flow of FIGS. 2 and 3 for the B-mode image of the two-layer sliding phantom of FIG. In FIG. 9A, it can be seen that the virtual image is suppressed at the lower part (deep part) of the image as in the case of p = 2 in FIG. 6B. However, a new virtual image is seen in the static region above the boundary line. This is presumed that the norm of p = 1 is susceptible to sparsity. Thus, in order to suppress the virtual image of the stationary part, it is desirable to set p to a value larger than 1. It is also possible to perform image processing for eliminating the virtual image. This will be described in the second embodiment.

  In the above description, the divergence degree is obtained as a statistic indicating the distribution state of the p-norm in the search region 32, and the scalar field image is generated from this value. However, other parameters other than the divergence degree can be used. . For example, a coefficient of variation can be used. The coefficient of variation is defined by the following equation, and is a statistic obtained by standardizing standard deviations on the average, and indicates the magnitude of distribution variation (that is, difficulty in separating minimum values).

(Second Embodiment)
In the second embodiment, when a virtual image is generated in the scalar field image obtained in the first embodiment, the virtual image is removed. That is, by identifying the reliability of the image area and removing the low reliability area, the virtual image is removed and the reliability of the entire image is improved. This will be described with reference to FIG. 9 and FIG.

  FIG. 9A is a scalar field image obtained by setting p = 1 in equation (1) as described in the first embodiment, and a virtual image is generated at the boundary. FIG. 9B is a histogram used for identifying the reliability of the image area, and FIG. 9C is a scalar field image in which the luminance of the image area with low reliability is replaced with a dark color. FIG. 10 is a flowchart showing the operation of the processing unit 10 for removing a virtual image.

  First, when receiving an instruction to remove a virtual image from the operator, the processing unit 10 operates as shown in the flow of FIG. 10 by reading and executing a program for removing a virtual image. First, one of the plurality of ROIs 31 set in step 51 of the flow of FIG. 3 is selected (step 102), and the p-norm value obtained for the plurality of candidate regions 33 in the search region 32 corresponding to this ROI 31 is processed. 10 is read from the memory 10b in the memory 10, the average is calculated, and the average value of the p-norm values corresponding to the ROI 31 is set (step 103). These steps 102 and 103 are repeated for all ROIs 31 (step 101).

  A histogram as shown in FIG. 9B is generated from the average value and the frequency of the p-norm values obtained for all the ROIs 31 (step 104). As the average value of the p-norm value is larger, the ROI is evaluated as having a lower reliability. If a low peak mountain distribution is generated in a region where the average value of the p-norm value is large in the histogram, this portion is less reliable. It is determined as the degree area 91. That is, the peak-like distribution having a large peak located in the region where the average value of the p-norm value is small is determined as the high-reliability portion 90, and the peak shape of the low peak located in the region where the average value of the p-norm value is larger than that Is determined to be the low reliability portion 91. Then, the position of the valley (frequency minimum value) 93 between the low reliability unit 91 and the high reliability unit 90 is obtained (step 105). The ROI 31 in the range where the average value of the p-norm value is larger than the valley 93 (low reliability region 91) is set as the low reliability region (step 106). A scalar field image is generated by removing the scalar values (deviation and variation coefficient) obtained in step 54 of FIG. 3 for the ROI 31 (step 107). For example, as shown in FIG. 9C, the ROI 31 in the low reliability region is displayed by replacing the luminance with a predetermined dark color. Further, the brightness of the ROI 31 in the low reliability area can be replaced with a predetermined bright color or displayed with the same brightness as the surrounding brightness.

  According to the second embodiment, since the virtual image can be removed, it is possible to provide a scalar field image in which the boundary of the subject can be recognized more clearly.

(Third embodiment)
In the first embodiment, the statistic of the p-norm distribution (deviation and variation coefficient) is obtained and an image is generated. However, in the third embodiment, the tissue boundary is determined from the p-norm distribution using another method. Generate a recognizable image. This processing method will be described with reference to FIGS.

  In the distribution of the p-norm value in the search area 32, the candidate area 33 along the boundary of the subject is an area having a small p-norm value (valley of p-norm value) along the boundary as described in the first embodiment. ) Is formed. For this reason, the distribution of the p-norm value is characterized in that the value of the candidate region 33 along the boundary shows a smaller value than the candidate region 33 along the direction orthogonal to the boundary. By utilizing this, an image is generated in the present embodiment.

  FIG. 11 shows a processing flow of the processing unit 10 of the present embodiment. 12A to 12H show eight patterns of candidate regions 33 selected on the p-norm distribution of the search region 32. FIG. FIGS. 12A to 12H show an example in which the candidate areas 33 are arranged in a 5 × 5 matrix in the search area 32 for easy illustration. The actual arrangement is the arrangement set in step 52.

  First, the processing unit 10 executes Steps 21 to 23 of FIG. 2 and Steps 51 to 53 of FIG. 3 of the first embodiment, and obtains a distribution of p-norm values for the search region 32 corresponding to a plurality of ROIs 31. Next, the ROI 31 is selected (step 111), and a predetermined direction 151 passing through the center of the search region 32 is set in the norm value distribution of the search region 32 corresponding to the ROI 31 as shown in FIG. For example, in the case of FIG. 12A, the predetermined direction 151 is a horizontal direction. A plurality of candidate areas 33 positioned along the set direction 151 are selected, and an average of norm values of these candidate areas 33 is obtained (step 114).

  The processes in steps 113 and 114 are all performed in the eight directions 151 shown in the eight patterns in FIGS. 12A to 12H (step 112). In the pattern of FIG. 12B, the predetermined direction 151 is a direction inclined about 30 ° counterclockwise with respect to the horizontal direction. In the patterns of FIGS. 12C to 12H, the predetermined directions 151 are about 45 °, about 60 °, 90 °, about 120 °, about 135 °, respectively, counterclockwise with respect to the horizontal direction. The direction is inclined at about 150 °.

  Among the eight predetermined directions 151, the direction 151 in which the average p-norm value is the minimum value is selected (step 115). Next, a direction 152 orthogonal to the selected direction 151 is set, and an average of the p-norm values of the candidate regions 33 located along the direction 152 is obtained (step 116). The direction 152 orthogonal to the eight directions 151 is as shown in FIGS. Ratio of the average of the p-norm values in the direction 151 where the average value of the p-norm values selected in step 115 is the minimum to the average of the p-norm values in the direction 152 orthogonal to the selected direction 151 determined in step 116 (= (Average of p-norm values in orthogonal direction 152 / average of p-norm values in minimum direction 151). This ratio is set as a pixel value (for example, luminance value) of the target ROI 31. An image is generated by executing this process for all the ROIs 31 (step 117).

  Since the candidate region along the subject boundary of the search region 32 has a small p-norm value (valley), the ratio obtained in step 117 is compared with the ROI 31 not located on the boundary in the ROI 31 located on the boundary. Larger value. Therefore, by setting the ratio as the pixel value, an image that can clearly recognize the boundary can be generated.

  In the present embodiment, the average ratio of the p-norm values is used. However, the present invention is not limited to this. Other functions such as the difference value between the average of the p-norm values in the minimum direction 151 and the average of the p-norm values in the orthogonal direction 152 are used. It is also possible to use a value.

In the above description, as shown in FIGS. 12A to 12H, the boundary of the subject is obtained from the valley of the distribution of the p-norm values of the candidate area 33 arranged in the search area 32. The present invention is not limited to this, and it is also possible to obtain the boundary of the subject from the distribution of pixel values in one candidate region 33 using the same method. Specifically, in FIGS. 12A to 12H, the search area 32 is replaced with a candidate area 33, and the candidate area 33 in the search area 32 is replaced with a pixel. In this case, in the example of FIGS. 12A to 12H, one candidate region 33 is composed of 5 × 5 pixels. In the candidate area 33, eight directions 151 passing through the central pixel of the candidate area 33 and directions 152 orthogonal to the directions 151 are set. The eight directions 151 and the eight directions 152 orthogonal thereto are each composed of five pixels. The pixel value of 5 pixels in each direction is P _{m + Δ} (i, j), the pixel value of the center pixel of 5 pixels is P _m (i ₀ , j ₀ ), and the equation (1) of the first embodiment Calculate the p-norm of 5 pixels in the direction. A p-norm average value obtained by dividing the obtained p-norm value by the number of pixels (5 in the case of 5 pixels) is obtained. The p-norm average value is obtained for each of the eight directions 151 in FIGS. 12A and 12B, and the direction 151 in which the p-norm average value is the minimum value is selected. The ratio of the p-norm average value in the direction 151 that is the minimum value and the direction 152 that is orthogonal to the direction 151 is calculated.

  In the candidate region 33 located at the boundary portion of the subject, the p-norm average value of the pixels in the direction along the boundary (the direction 151 in which the p-norm average value is minimum) is small, and the p-norm average value in the orthogonal direction 152 is Since it becomes large, the ratio of both becomes a large value. On the other hand, in the candidate region 33 located in a uniform region other than the boundary of the subject, the p-norm average value in the direction 151 and the p-norm average value in the orthogonal direction 152 are equal, so the ratio between them is 1 Close to. Therefore, if the ratio is calculated for the candidate area 33 of the entire image of the target frame, the pixel of the candidate area 33 with a large ratio corresponds to the boundary portion. By generating the image, it is possible to generate an image in which the boundary line can be estimated in units of pixels. Instead of the ratio, it is also possible to use other function values such as a difference value between the p-norm average value in the minimum direction 151 and the p-norm average value in the orthogonal direction 152.

(Fourth embodiment)
A fourth embodiment will be described.

  In the fourth embodiment, the processing unit 10 applies the Laplacian filter to the p-norm distribution and emphasizes the p-norm distribution before performing the processing of FIG. 11 on the p-norm distribution of the search area 32 in the third embodiment. Apply processing. Since the valley of the p-norm value along the boundary direction is emphasized by performing the enhancement process on the p-norm distribution, the ratio value obtained by performing the process of FIG. It is possible to obtain an image with a large contrast between the portion and the region that is not.

  Specifically, Steps 21 to 23 in FIG. 2 and Steps 51 to 53 in FIG. 3 of the first embodiment are executed to obtain a distribution of p-norm values for the search region 32 corresponding to a plurality of ROIs 31. A spatial second-order differential process (Laplacian filter) is applied to the obtained distribution of p-norm values to generate a p-norm value distribution in which the contours of the valleys of the p-norm values along the boundary direction are emphasized. The p-norm distribution after applying the Laplacian filter is subjected to the process of FIG. 11 of the third embodiment to generate an image.

  Similarly, when obtaining the boundary of the subject from the distribution of pixel values in the candidate area 33 described in the second half of the third embodiment, after applying a Laplacian filter to the distribution of pixel values and applying enhancement processing, It is also possible to obtain the p-norm average value and ratio.

(Fifth embodiment)
As a fifth embodiment, a processing method for generating an image capable of recognizing a tissue boundary from a p-norm distribution using eigenvalue decomposition processing will be described.

First, the processing unit 10 executes Steps 21 to 23 of FIG. 2 and Steps 51 to 53 of FIG. 3 of the first embodiment, and obtains a distribution of p-norm values for the search region 32 corresponding to a plurality of ROIs 31. A matrix A is generated using the p-norm value (N _mn ) of the candidate area 33 in the obtained search area 32. The matrix A is substituted to a unique equation of the following formula (4), the eigenvalues λ _n, λ _n-1, obtaining a · · · lambda _1. Among the eigenvalues, a maximum eigenvalue or a linear combination of eigenvalues is set as a scalar value of the ROI 31 corresponding to the search region 32. Here, the linear combination of eigenvalues means, for example, using two of the maximum eigenvalue λn and the second eigenvalue λ _n−1 and using the function, for example, λ _n −λ _n−1 as a scalar value. .

ただし、N_mnは、探索領域３２内の候補領域３３について式（１）により求めたｐノルム値であり、ｍおよびｎは、探索領域３２内の候補領域３３の位置を示す。

Here, N _mn is the p-norm value obtained by the equation (1) for the candidate area 33 in the search area 32, and m and n indicate the position of the candidate area 33 in the search area 32.

  For all ROIs 31, the maximum eigenvalue or a linear combination of eigenvalues is obtained as a scalar value, and a scalar field image having a scalar value as a pixel value (such as a luminance value) is generated in the same manner as in step 56 of FIG.

  Thus, in the present invention, a scalar field image can be generated using eigenvalues.

  In the present embodiment, the maximum eigenvalue among eigenvalues or a linear combination of eigenvalues is used. However, the present invention is not limited to this, and one or more other eigenvalues can be used.

(Sixth embodiment)
As a sixth embodiment, a processing method for generating a scalar field image from which a boundary can be extracted based on a vector field when step 25 of FIG. 8 is performed to generate a motion vector field image in the first embodiment. explain.

  It is assumed that the motion vector field obtained in step 25 of FIG. 8 is a model as shown in FIGS. 13 (a), (b), and (c). The model of FIG. 13A is an example of the horizontal direction through the ROI (target pixel) 131 whose center is the boundary direction of the subject, and the model of FIG. For example, the model in FIG. 13C is an example in which the direction of the boundary is an oblique 45 degree direction. The subject assumes that the region across the boundary is moving with a motion vector of size c in the opposite direction.

  First, a case where the vector field of such a model is converted to a scalar field by obtaining a conventional distortion tensor will be described. The mathematical formula for obtaining the strain tensor is a well-known formula described in Patent Document 2, and is defined by the following formula.

式（５）において、動きベクトルのｘ成分をX、ｙ成分をYとしている。

In equation (5), the x component of the motion vector is X and the y component is Y.

  The partial differential value represented by Expression (5) can be calculated, for example, as a difference average of vector components on both sides of the ROI 131. Specifically, equation (6) is obtained for each model in FIGS. 13 (a), (b), and (c).

  For example, (0, C) in the vector field of FIG. 13 (a), (−C, 0) in the vector field of FIG. 13 (b), (−C / √2, in the vector field of FIG. 13 (c). C / √2). For this reason, even if the vector field on both sides across the boundary has the same condition that the magnitude is C and the direction is opposite, if the boundary is oblique as shown in FIG. Since the tensor becomes C / 2 of the distortion tensor in the case of a horizontal or vertical boundary as shown in FIGS. 13A and 13B, the vector field is converted into a scalar field by the distortion tensor, and the distortion tensor is converted. Is imaged as a pixel value of the ROI 131, the boundary in the oblique direction becomes unclear compared to a horizontal or vertical image. For this reason, the detection ability of the boundary of an oblique direction falls.

  Therefore, in the present invention, a motion vector field is converted into a scalar field using a scalar value defined by the following equation (7). Since Expression (7) is a form including a power and a power root similar to Expression (1), it is referred to as a boundary norm here.

  In each model of FIGS. 13A, 13B, and 13C, when the boundary norm is obtained, the boundary norm value is C in any model. Thus, the vector field can be converted equally to a scalar field regardless of the direction of the vector. Therefore, by using the boundary norm of the present invention, a vector field is converted into a scalar field, and an image having a scalar value (boundary norm value) as a pixel value of the ROI 131 can be generated to detect a boundary having high robustness with respect to the direction. It becomes.

  FIG. 14 shows a procedure for generating a scalar field image using the boundary norm of this embodiment. First, steps 21 to 25 in FIG. 8 of the first embodiment are executed to generate a vector field image. The process of FIG. 14 is performed with respect to the generated vector field image. First, a plurality of ROIs 131 are set on the vector field image (step 142). Next, partial differentiation of vectors is performed for one ROI 131 in the x direction and y direction, and the boundary norm of Expression (7) is calculated using this (step 143). The obtained boundary norm value is set as the scalar value of the ROI 131. These processes are repeated for all ROIs 131 (step 141). Then, the boundary norm value is converted into a pixel value (for example, luminance value) of the ROI 131 to generate a scalar value image.

(Seventh embodiment)
A seventh embodiment will be described.

  In the seventh embodiment, when a plurality of ROIs 31 are set in step 51 of FIG. 3 of the first embodiment, when the ROIs 31 are arranged so as to overlap each other for improving resolution as shown in FIG. In step 53, the calculation result of the overlapping area 151 is stored in a lookup table provided in the memory 10b of the processing unit 10 to reduce the calculation amount. Other configurations are the same as those of the first embodiment.

  FIG. 16 shows the processing procedure of step 53 in the present embodiment. In the flow of FIG. 16, the same steps as those in the flow of FIG. 3 are denoted by the same reference numerals.

  First, when there are overlapping areas 151-1 and 151-2 in the plurality of ROIs 31 set in step 51 of FIG. 2 as shown in FIG. 15, the p-norm sum of Expression (8) described later for the pixels in the overlapping area Is registered in the memory 10b of the processing unit 10 to generate a lookup table (step 161).

  Next, the target ROI 31-1 is selected (step 163), and the candidate area 33 in the search area 32 corresponding to the target ROI 31-1 is selected. Of the pixels of ROI 31-1, for which the p-norm sum is not stored in the lookup table (that is, pixels that are not the overlapping region 151-1), the following equation (1) is an equation in the p-power root of (1): According to 8), the p-norm sum is calculated (step 165). In step 165, if the p-norm sum data of the overlapping area 151-1 is not recorded in the lookup table, the p-norm sum is also calculated for the pixels of the overlapping area 151-1.

  Next, referring to the look-up table, if the p-norm sum data of the pixel in the overlap region 151-1 of the ROI 31-1 is stored, it is read out and added to the p-norm sum obtained in step 165. Then, the p-norm of the equation (1) is obtained by calculating the p-th power root of the addition result (step 166). Thereby, the p-norm value for the candidate region of the ROI 31-1 can be obtained. The obtained p-norm value is stored in the memory 10b.

  If the p-norm sum calculated in step 166 includes the data of the unrecorded overlapping area 151-1 in the lookup table, the p-norm sum of the overlapping area 151-1 is recorded in the lookup table (step 167). It repeats about all the candidate area | regions in the search area | region 32 corresponding to ROI31-1. Thereby, the p-norm value distribution of ROI 31-1 can be obtained (step 168). If the p-norm value distribution for the ROI 31-1 is obtained, the degree of divergence is obtained in step 54 and set as the scalar value of the target ROI 31-1.

  Next, the next ROI 31-2 is selected (steps 162 and 163), and a candidate area is selected (step 164). Among the pixels of ROI 31-2, for the pixels in which the p-norm sum is not stored in the lookup table (that is, pixels that are not the overlapping region 151-1), the following equation (1) is an equation in the p-power root of (1): According to 8), the p-norm sum is calculated (step 165). Since the p-norm sum data of the pixels in the overlap region 151-1 of the ROI 31-2 is already stored in the lookup table, it is read out and added to the p-norm sum obtained in step 165, and the addition result p The p-norm of equation (1) is obtained by calculating the power root (step 166). Thus, the p-norm value for the candidate region of the ROI 31-2 can be obtained with a small amount of calculation without calculating the p-norm sum of the overlapping region 151-1.

  The obtained p-norm value is stored in the memory 10b. In addition, the p-norm sum of the overlapping area 151-2 obtained in the calculation in step 165 is recorded in the lookup table (step 167).

  The p-norm value distribution can be obtained by repeating the above steps 163 to 168 for all ROIs (step 55). This eliminates the need for recalculation in the overlapping area 151, and reduces the amount of calculation.

  In the present embodiment, the configuration in which the overlapping region is set and the p-norm sum is stored in the lookup table when the adjacent ROIs 31 partially overlap has been described. Even when the region 33 partially overlaps, the amount of calculation can be reduced by setting an overlapping region and storing the p-norm sum of the region in the lookup table.

(Eighth embodiment)
An eighth embodiment will be described.

  By executing any one of the first to seventh embodiments described above for a continuous frame, a continuous image of a scalar field or a continuous image of a vector field generated from a norm distribution can be generated. Can be displayed in series. At this time, an abnormal frame that has not been properly generated for some reason may occur. In the eighth embodiment, abnormal frames are removed and an appropriate continuous image can be displayed.

ここで、Pxは、ベクトルのｘ成分の発生確率、Pyは、ベクトルのｙ成分の発生確率である。この式により求められる情報エントロピーHは、x成分とｙ成分の結合エントロピーであり、フレーム全体の平均情報量を表している。The abnormal frame has a feature that the drawing area is extremely small. Therefore, the abnormal frame is determined to determine whether it is an abnormal frame or a normal frame. In the present embodiment, whether the rendering area is large or small is determined based on the magnitude of information entropy. The information entropy of the vector field image is defined by the following equation (9).

Here, Px is the occurrence probability of the x component of the vector, and Py is the occurrence probability of the y component of the vector. The information entropy H obtained by this equation is the combined entropy of the x component and the y component, and represents the average information amount of the entire frame.

  When the information entropy is obtained for the scalar field image obtained from the p-norm distribution or the like according to the first to seventh embodiments, the variable is only one item on the right side of Expression (9).

  FIG. 17 shows a result of obtaining information entropy by the above equation (9) for an example of time-series continuous frames. FIG. 18 shows the change in information entropy over time for 10 consecutive frames. FIG. 17 shows an image frame display processing procedure of the present embodiment. Since a frame with a small information entropy is an abnormal frame with a small amount of information, a frame in which the information entropy is less than a predetermined threshold is not displayed, and a process of holding and displaying the previous frame is performed.

  Specifically, a threshold is set in step 181 in FIG. 18, the first frame is selected, information entropy is calculated, and if it is less than the set threshold, the current frame (the current frame) in which information entropy is calculated Instead, the previous frame is displayed, and if it is equal to or greater than the threshold, the current frame is displayed as it is. This is repeated for all frames. With this process, it is possible to remove abnormal frames and display a continuous image with good visibility. As the threshold value, for example, a predetermined value can be used, or an average value of a predetermined number of frames can be used.

(Ninth embodiment)
A ninth embodiment will be described.

  In the first embodiment, a scalar value image obtained from a p-norm distribution and a B-mode image are superimposed and displayed as shown in FIG. 7A, and in FIG. 7B, a vector field is further added to this image. The image is superimposed and displayed. In the ninth embodiment, visibility is improved by extracting only the boundary portion of the scalar field image to be superimposed as shown in FIG. 19A and superimposing it on the B-mode image or the like.

  FIG. 20 shows an image composition processing procedure according to this embodiment. First, a histogram of scalar values of a scalar field image generated in the first embodiment or the like is created as shown in FIG. 19B (step 201). A mountain-shaped distribution in a region having a large scalar value is searched, and its minimum value (distribution trough) 191 is searched (step 202). Then, the minimum value 191 is set as a threshold, pixels having a scalar value larger than that are extracted from the scalar field image, and an extracted scalar field image is generated (step 203). The extracted scalar field image depicts a boundary portion having a large scalar value. By displaying the extracted scalar field image superimposed on the B-mode image (and the vector field image), the boundary can be clearly recognized, and the area other than the boundary can be easily confirmed with the B-mode image and the vector field image. An image with high visibility can be displayed (step 204).

  INDUSTRIAL APPLICABILITY The present invention can be applied to medical ultrasonic diagnostic / treatment apparatuses and general apparatuses that measure distortion and deviation using general electromagnetic waves including ultrasonic waves.

  1: Ultrasonic probe (probe) 2: User interface 3: Transmission beamformer 4: Control system 5: Transmission / reception changeover switch 6: Reception beamformer 7: Envelope detector 8: Scan converter, 10: processing unit, 10a: CPU, 10b: memory, 11: parameter setting unit, 12: synthesis unit, 13: display unit

Reception beam former 6 performs dynamic focus adjusting delay time according to reception timing under the control of the control system 4 and outputs the RF signal of a real part and an imaginary part. The RF signal is detected by the envelope detector 7 and then converted into a video signal, which is input to the scan converter 8 and converted into image data (B-mode image data). The configuration described above is the same as the configuration of a known ultrasonic imaging apparatus. Furthermore, in the present invention, it is also possible to perform ultrasonic boundary detection by a configuration that directly processes the RF signal.

As a comparative example, FIG. 6C shows a vector field image obtained by a conventional block matching process. In this image, the B-mode image in FIG. 6A is set as the frame m, the ROI moving position is obtained by block matching processing (step 24) with the next frame (frame m + Δ, Δ = 1 frame), and the moving direction is determined. It is the vector field image which showed the magnitude | size as a vector (arrow). In FIG. 6C, a phenomenon in which the motion vector is disturbed is seen in the lower (deep) region. This is because the SN ratio of the detection sensitivity decreases and the influence of electric noise and the like increases as the distance from the probe installed at the upper portion increases, indicating the penetration limit. FIG. 6D shows a distortion tensor of the vector field shown in FIG. 6C, which is imaged as a pixel value (for example, a luminance value). In FIG. 6 ( d ), a phenomenon in which a virtual image is generated in the deep part due to the influence of the erroneous vector can be seen.

As described above, the scalar field image (FIG. 6B) obtained from the p-norm of the present embodiment is compared with the conventional vector field of FIG. 6C and FIG. 6D and the distortion tensor field image based on the vector field. Thus, the virtual image can be suppressed and the slip boundary portion can be clearly shown.

First, when there are overlapping areas 151-1 and 151-2 as shown in FIG. 15 in the plurality of ROIs 31 set in step 51 of FIG. 3 , the p-norm sum of the later-described equation (8) for the pixels in the overlapping area Is registered in the memory 10b of the processing unit 10 to generate a lookup table (step 161).

Next, select the following ROI31-2 (Step 163), selects the candidate area (step 164). Among the pixels of ROI 31-2, for the pixels in which the p-norm sum is not stored in the lookup table (that is, pixels that are not the overlapping region 151-1), the following equation (1) is an equation in the p-power root of (1): According to 8), the p-norm sum is calculated (step 165). Since the p-norm sum data of the pixels in the overlap region 151-1 of the ROI 31-2 is already stored in the lookup table, it is read out and added to the p-norm sum obtained in step 165, and the addition result p The p-norm of equation (1) is obtained by calculating the power root (step 166). Thus, the p-norm value for the candidate region of the ROI 31-2 can be obtained with a small amount of calculation without calculating the p-norm sum of the overlapping region 151-1.

FIG. 17 shows a result of obtaining information entropy by the above equation (9) for an example of time-series continuous frames. FIG. 17 shows the change over time in information entropy for 10 consecutive frames. FIG. 18 shows an image frame display processing procedure of the present embodiment. Since a frame with a small information entropy is an abnormal frame with a small amount of information, a frame in which the information entropy is less than a predetermined threshold is not displayed, and a process of holding and displaying the previous frame is performed.

Claims (20)

A transmission unit that transmits ultrasonic waves toward the target; a reception unit that receives ultrasonic waves coming from the target; and a processing unit that generates an image of two or more frames by processing a reception signal of the reception unit. And
The processing unit sets a plurality of regions of interest in one frame among the generated images of two or more frames, and sets a search region wider than the region of interest in another frame for each of the plurality of regions of interest. And setting a plurality of candidate regions having a size corresponding to the region of interest in the search region, and setting a norm between the pixel value of the region of interest and the pixel value in the candidate region. And calculating a norm distribution in the search region and generating a value representing a distribution state of the norm as a pixel value of the region of interest corresponding to the search region. Acoustic imaging device. 2. The ultrasonic imaging apparatus according to claim 1, wherein the norm is a p-norm represented by the following formula (1).

However, P _m (i ₀ , j ₀ ) is the pixel value of the pixel at the position (i ₀ , j ₀ ) in the region of interest, and P _{m + Δ} (i, j) is the position (i , J) The pixel value of the pixel in p), p is a predetermined real number   The ultrasonic imaging apparatus according to claim 2, wherein the p is a real number larger than one.   The ultrasound imaging apparatus according to claim 1, wherein the value representing the norm distribution state is a statistic of the distribution.   5. The ultrasonic imaging apparatus according to claim 4, wherein the statistic is a divergence defined by a difference between a minimum value of a norm value of a norm distribution in the search region and an average value of the norm values. A featured ultrasonic imaging apparatus.   5. The ultrasonic imaging apparatus according to claim 4, wherein the statistic is a coefficient of variation obtained by dividing a standard deviation of a norm value of a norm distribution in the search region by an average value.   The ultrasound imaging apparatus according to claim 1, wherein the processing unit is a norm value of the candidate region at a position along the direction among a plurality of directions centered on the region of interest set in the search region. A first direction that minimizes an average of the second direction and a second direction that is orthogonal to the first direction through the region of interest, and an average of the norm values of the candidate regions along the first direction; A ratio value or a difference value with respect to an average of the norm values of the candidate regions along the second direction is used as a value representing a distribution state of the norm for the region of interest corresponding to the search region. An ultrasonic imaging apparatus.   2. The ultrasound imaging apparatus according to claim 1, wherein the processing unit sets a central pixel at a center in the candidate region, sets a plurality of directions around the central pixel, and sets a pixel value of the central pixel. And a norm value between the pixel values of the plurality of pixels at the position in the direction, an average of norm values obtained by dividing the obtained norm value by the number of pixels in the direction is obtained, and the average of the norm values is minimized. A ratio value or a difference value between an average of norm values obtained for one center and a plurality of pixels located in a second direction orthogonal to the first direction through the center pixel and the center pixel is calculated as the candidate area. An ultrasonic imaging apparatus characterized by being used as a value of the center pixel of the image.   The ultrasound imaging apparatus according to claim 7, wherein the processing unit performs an enhancement process on a norm distribution in the search region in advance by a Laplacian filter, and obtains the ratio value or the difference value for the distribution after the enhancement process. A featured ultrasonic imaging apparatus.   9. The ultrasonic imaging apparatus according to claim 8, wherein the processing unit performs an enhancement process on a pixel value in the candidate area in advance by a Laplacian filter, and obtains the ratio value or the difference value for the pixel value after the enhancement process. A featured ultrasonic imaging apparatus.   The ultrasound imaging apparatus according to claim 1, wherein the processing unit generates a matrix representing a norm distribution in the search region, performs eigenvalue decomposition processing on the matrix to obtain an eigenvalue, An ultrasound imaging apparatus, characterized in that it is used as a value representing a distribution state of the norm for the region of interest corresponding to a search region. The ultrasound imaging apparatus according to claim 1,
The processor is
Selecting a candidate region having the smallest norm value in the search region as a destination of the region of interest, obtaining a motion vector connecting the position of the region of interest and the position of the selected candidate region, Generating a motion vector field by generating the motion vector for each region of interest;
For each of the plurality of regions of interest set in the motion vector field, the sum of the square value of the y-direction derivative with respect to the x component and the square value of the x-direction derivative with respect to the y component is obtained as a boundary norm value. An ultrasonic imaging apparatus that generates an image using a norm value as a pixel value of the region of interest. The ultrasound imaging apparatus according to claim 1,
The processing unit sets the plurality of regions of interest to partially overlap each other, and stores a value obtained for the overlapping regions when calculating a norm for one region of interest in a lookup table of a storage region. An ultrasonic imaging apparatus that is used by reading from the lookup table when calculating a norm for another region of interest. The ultrasound imaging apparatus according to claim 1,
The processing unit sets the plurality of candidate areas so as to partially overlap, and stores a value obtained for the overlapping areas when calculating a norm for one candidate area in a lookup table of a storage area. An ultrasonic imaging apparatus that is read from the lookup table and used when calculating norms for other candidate regions. The ultrasound imaging apparatus according to claim 1,
The processor is
An image generated by using a value representing the norm distribution state as a pixel value is generated in a plurality of frames in time series, and an information entropy amount is calculated for each frame,
When the information entropy amount is equal to or greater than a preset threshold value, the frame is displayed. The ultrasound imaging apparatus according to claim 1,
The processor is
An ultrasonic imaging apparatus, wherein an extracted image obtained by extracting pixels having a value representing a norm distribution state equal to or greater than a predetermined value is generated and superimposed on a B-mode image. The ultrasonic imaging apparatus according to claim 16, wherein
The processor is
For an image generated using a value representing the norm distribution state as a pixel value, a histogram representing the norm distribution state and its frequency is generated.
An ultrasonic imaging apparatus characterized by searching for a mountain-shaped distribution of the histogram and using a minimum value of the mountain-shaped distribution as the predetermined value. Transmitting an ultrasonic wave toward the object, processing the reception signal obtained by receiving the ultrasonic wave coming from the object, and generating an image of two or more frames,
Select 2 frames from the image,
Set multiple regions of interest in one frame,
A search area wider than the region of interest is set for each of the plurality of regions of interest in another frame, and a plurality of candidate regions having a size corresponding to the region of interest are set in the search region,
By obtaining a norm between the pixel value of the region of interest and the pixel value in the candidate region for each of the plurality of candidate regions, a distribution of norms in the search region is obtained,
An ultrasonic imaging method, wherein an image is generated using a value representing a distribution state of the norm as a pixel value of the region of interest corresponding to the search region. On the computer,
A first step of selecting two frames from two or more ultrasound images;
A second step of setting a plurality of regions of interest in one frame;
A third step of setting a search region wider than the region of interest in another frame for each of the plurality of regions of interest, and setting a plurality of candidate regions of a size corresponding to the region of interest in the search region;
A fourth step of determining a norm distribution in the search region by determining a norm between the pixel value of the region of interest and the pixel value in the candidate region for each of the plurality of candidate regions;
A fifth step of generating an image using a value representing the distribution state of the norm as a pixel value of the region of interest corresponding to the search region;
Ultrasound imaging program for running A transmission unit that transmits ultrasonic waves toward the target; a reception unit that receives ultrasonic waves coming from the target; and a processing unit that generates an image of two or more frames by processing a reception signal of the reception unit. And
The processing unit sets a plurality of regions of interest in a received signal distribution corresponding to one frame among received signals corresponding to the received image of two or more frames, and a received signal distribution corresponding to another one frame A search area wider than the region of interest is set for each of the plurality of regions of interest, a plurality of candidate regions having a size corresponding to the region of interest are set in the search region, and the amplitude distribution or phase distribution of the region of interest And the norm between the amplitude distribution or the phase distribution in the candidate region for each of the plurality of candidate regions, the norm distribution in the search region is obtained, and a value representing the distribution state of the norm is obtained. An ultrasonic imaging apparatus that generates an image as a pixel value of the region of interest corresponding to the search region.

Images